jazzpush2

5 ms·

1.
by jazzpush2 6d ago
Nothing new here in the idea, implementation, or narrative. Someone just pointed to an AI and said, 'Create a website explaining LLM's step-by-step.' Nothing is even visualized.
2.
by jazzpush2 7d ago
These social science/econ style jobs at frontier labs are the ultimate free ride. Get paid bank with absolutely zero need to publish anything of quality.
3.
by jazzpush2 7d ago
I would legitimately purchase a 'smart' phone that's just basically a 15-year old flip phone with maps and a modern battery + software stack. Awaiting the inevitable analog trend to kick back up so there are solid options her
4.
by jazzpush2 8d ago
I'm responding to a post, citing a particular research method, which the author claims six times to display the model's thinking. You can dismiss the criticism with pedantics about what he means by 'thinking', but what I
5.
by jazzpush2 8d ago
Zuckerberg: "Yeah so if you ever need info about anyone at Harvard. Just ask. I have over 4,000 emails, pictures, addresses, SMS" Friend: "What? How'd you manage that one?" Zuckerberg: "People just submitted it
6.
by jazzpush2 8d ago
" the intermediate activations of an LLM to decode what it is most likely going to say or is thinking about." There is no thinking in these models. The J space is a basic technique measuring how much the influence of shifting a to
7.
by jazzpush2 8d ago
Can you say anything specific? If you're going to criticize policy, there are many specific things you can say. 'Keeps China down' is not only incorrect, but meaningless itself.
8.
by jazzpush2 8d ago
I know someone who shared this sentiment. As if being highly-supported by the largest 'alt' media (Rogan etc.) and traditional media (Fox, etc.) is in any way rebellious. Never mind being a literal billionarie who got away with fu
9.
by jazzpush2 14d ago
You present the false dichotomy without qualifying 'the same way'. Is it in terms of collective taxes paid? Does that include or exclude family? Or is it in terms of money spent in the local economy? Or in terms of money sent abro
10.
by jazzpush2 14d ago
O-1 is about as far as you can get these days. Just search O-1 on x.com and see the type of people that have it - often designers, 'product' folks, etc. working for no-name YC startups.
11.
by jazzpush2 14d ago
I have multiple friends who studied CS and graduated in the last few years and couldn't find a job. These are people who grew up in the states, paid taxes here, came from (sometimes generations of) families who paid taxes here. I mysel
12.
by jazzpush2 15d ago
I am a researcher. Researchers, now more than ever with AI, have plenty of time to write something. The lift of writing is never the hard part, anyways.
13.
by jazzpush2 15d ago
Shame, it's a cool idea, but the writing is so full of claudism's my mind skims over it completely. Bit of an embarrassment to publish that as a published researcher, to be honest.
14.
by jazzpush2 16d ago
This is the best thing I've read in years.
15.
by jazzpush2 16d ago
That's my point. It's not a prerequisite for them, so they're able to get accepted with remdial math scores, making the entire university look bad.
16.
by jazzpush2 17d ago
With enough researcher degrees of freedom, you can make the data appear to support almost any reasonable story, sure. But that's not what's going on here - unless you have a pointed issue with the methodology, which if so, please
17.
by jazzpush2 17d ago
This is basically exclusively for the Humanities students at Cal. Majority of the STEM ones all came in with the calculus sequence finished already. Another big issue is that many of these under-prepared students came from the same Bay Area
18.
by jazzpush2 17d ago
There is still yet no reckoning for Zuckerberg nor the people who choose to work for Meta. What would justice look like for someone responsible for building something so detrimental to society? What of the employees?
19.
by jazzpush2 18d ago
To be fair, you were the one complaining about water costs as if it was a climate-related issue, when it's... your pool.
20.
by jazzpush2 21d ago
NVIDIA had more open-source models on huggingface than any other company. A lot better them than any other frontier lab, tbh.
21.
by jazzpush2 21d ago
Pure AI slop article.
22.
by jazzpush2 21d ago
Shame on anyone working there in 2026.
23.
by jazzpush2 21d ago
Absolutely insane videos coming out. I actually thought some were AI-generated when I first saw them.
24.
by jazzpush2 21d ago
It was certainly almost RL-fried to overfit the benchmarks, at the expense of actual usability. See Opus 5.
25.
by jazzpush2 22d ago
"all the issues" - which issues exactly? It didn't explain any of the contributing issues at all - it was a think piece on motivations for lack of action. To answer your question: literally every linked YouTube video.
26.
by jazzpush2 22d ago
Just let Perplexity die already.
27.
by jazzpush2 22d ago
You can review my post history to see I'm not astroturfing. I don't know what else to say except continue coping. Try searching 'opus 5' on twitter to see what a shit show that model is.
28.
by jazzpush2 22d ago
I have completed stopped using Anthropic since Kimi 3 released, and only ever K3 and Codex nowadays. I pay for K3, and get Codex and Claude free from work. Almost everyone I know prefers Codex to Claude.
29.
by jazzpush2 23d ago
This was a fun, light read, but nowhere near 'best', hardly more than a fun opinion piece, and even then it completely ignored relevant topics (class inequality at foot, performative climate action [taylor swift], disparate impact
30.
by jazzpush2 29d ago
You really can't think of any current events that would make this top-of-mind? Nothing over the last ~2 years?

More ›